A Soft-roadbed Settlement Prediction Model Based on RBFNN

摘要

The control of subgrade settlement has been the main influencing factor on highroad pavement quality, so estimating settlement amount after construction based on designed embankment height and planning the construction filling height beforehand through the settlement prediction is a very important work. Because of the complex coupling properties and high nonlinear characteristics of the factors influencing foundation settlement, a soft-roadbed settlement prediction model based on RBFNN (Radial Basis Function Neural Network) is proposed, combining the global fitting properties of RBFNN. The analysis shows that the model owns good fitting property and high prediction precise, and certain practical value.